It says its using the GPU but if you look at the cpu time required to play a 1080p video within firefox vs playing the same video with mpv
www.mpv.io : a lightweight player that among other things can transparently stream any url understood by youtube-dl
It uses 20x as much cpu as mpv. For purposes of comparison this is like Toyota selling a car that goes just as fast as your Ford so long as you are OK with only getting 4MPG.
There is an addon if you prefer to right click on a url and open in mpv or in an addon like Tridactyl you can do the same with your keyboard. It also lets you speed up and slow down the playback speed which is nice for speakers who talk too slow and take a while to get to the point.
Given that Linux has 2% desktop market share (let's be generous) and Firefox has 10% (let's be generous the same way), that would mean that Linux has potential to be 20% of Firefox users. Is that something you are going to ignore?
Many executives would sell their families for less opportunity.
Linux distros tend to ship with firefox and linux users are more concerned with open source. It would be strange if the percentage of mozillas userbase using linux was smaller than the general percentage of pc users running linux.
Its probably more fair to go on profit not revenue.
100 million in revenue * 2.18% is 2.2 million. A decrease of 10% of that is 220k. If engineers at mozilla are 150k then 3 months of work is aprox 38k.
This would put the break even point at losing 2% of its linux users by not spending the money.
according to whom? this will almost certainly need more than a single engineer's eyes on it. and even if it is 3 months, that's 3 months they don't spend working on features for 98% of their userbase.
i want this to happen as much as anyone, but i just don't see the compelling math here. it's a chicken-egg problem (if FF could use VA API [1], then more might switch to linux).
personally, i'm happy that Panfrost recently got mainlined.
